OBJECTIVE: Evaluation of the effectiveness of third trimester fetal biometry and Doppler studies in predicting the birth of a large for gestational age neonate (LGA ≥ 95 centile). Assessment of the value of integrated models (combining maternal characteristics, first trimester parameters, third trimester fetal biometry) and the usefulness of contingency strategies. METHOD: Observational cross-sectional study on 2308 uncomplicated singleton pregnancies examined at 11 to 14 weeks and at 30 to 34 weeks. RESULTS: Ultrasound estimated fetal weight (EFW, area under the curve (AUC) = 0.83) was the best single predictor of LGA. Maternal weight, delta nuchal translucency and EFW were independent predictors for the integrated model, but the latter was not statistically better (AUC = 0.84) than using EFW alone. The detection rates for LGA were 72.5% and 73.7% for a 25% screen positive rate, by EFW and the third trimester integrated model respectively.A contingency strategy of rescanning 50% of the population in the third trimester according to the risk estimation by a first trimester prediction model results in detection rate of 64.7% for LGA for the same 25% screen positive rate (AUC = 0.78). CONCLUSIONS: Third trimester ultrasound is an effective screening modality for identifying fetal macrosomia. A contingency strategy utilizing first trimester parameters can reduce the need for unnecessary examinations. © 2012 John Wiley & Sons, Ltd.